5973 Strange Issue

[WahWash!]

Hello!

We're a small company who bought a decommissioned 6890/5973N from auction a few months ago. After scouring manuals, buying consumables, and a lot of reading we were able to get the instrument up and running a few weeks ago with beautiful clean results!

Until today.

Went to fire it up and pump down and noticed some big unusual jumps on the foreline pressure in the first few minutes, but it stabilized at a healthy 50 mTorr and we let it sit for a few hours before attempting a tune. As soon as we started a tune I got a message saying Log Amp Mismatch and it stopped tuning.

Now every time I attempt to start the MSD it thinks it has a turbo pump (it doesn't) and shows turbo pump speed (0%) instead of foreline pressure.
All heaters work just fine and can be controlled remotely or on the panel, and the rough pump runs just fine. The rest of the panel interface works like normal.

We've checked every fuse, including the ones that require removing the board enclosure. All wires and connectors were removed and reseated with no luck.

I see no evidence of the diffusion pump kicking on (no red LED)

Any ideas? Could a corrupted firmware possibly cause something like this?
(Update: Flashing firmware didn't fix it)

Thanks for reading!

[WahWash!]

Oh and the weirdest things:

The MSD Version screen shows no serial number and
The LOG Amplifier shows LOG00,119,51,119,3AL

[70 eV]

Is your mass analyzer door creating a proper seal? I know the diffusion pumps needed oil, I've found old bottles around the lab from decades ago.

I'd keep an eye out on govt auction websites tbh, you can score a 5975 for under 1k with freight included if you shop around enough. Our lone 5975 hasn't been ran in 3 years.

[benhutcherson]

I'd probably try taking the hi-vac control cable-the one with the Db-25 connector that runs to the convectron gauge, the pump temperature sensors, and the front fan and checking for continuity with VOM. I don't have a pinout for it, but you should be able to match every connection to a pin on the connector.

With that said, you have errors all over the place. Both the sideboard and the log amp came in "slow" versions(first model 5973s up until I think almost the end of HP badging) and "fast" versions(any 73 that was originally Agilent badged I THINK should be this). I don't know exact part numbers, but know that the slow sideboard starts with G1099 and I'd not be surprised if the log amp starts with the same. The log amp mismatch error usually happens when you change one without changing the other.

The switch to thinking it has a turbo is weird. I actually relatively recently changed my 5973 from diffusion to turbo, or actually more properly I bought an untested turbo instrument off Ebay and transplanted parts from my diff instrument to it(the one I bought was missing a transfer line and ion source plus was HPIB where mine has been upgraded to LAN-I dropped my complete sideplate/analyzer onto it and moved the transfer line over). There is no difference I could find between a turbo and diff pump emod-they have the same PN and all boards are the same PN. The power board(the one where the wall power comes in) does have a jumper that says "diffusion pump power" but it was in the same position on both units. Where I'm really going with this, though, is that Chemstation had only ever known this as a diff pump instrument, but the first time I powered it up with the emod from the diff pump instrument(I'd tested it a fair bit with the HPIB emod it shipped with, but didn't want that long term), Chemstatoin had zero issues picking up that it was a turbo instrument and just immediately switched to reporting turbo speed without even a popup saying "It's different now."

Where I'm really going with this, though, is that the errors that-in my minds are unconnected(log amp+vac pump)-makes me think you have a board issue. The question for me would be which board. My mind actually goes to the SmartCard since it's the one that handles all the com/IO and actually more or less the "brains" of what's going on in the MS, but it could also be the main board.

[benhutcherson]

Is your mass analyzer door creating a proper seal? I know the diffusion pumps needed oil, I've found old bottles around the lab from decades ago.

Oil is kind of how diff pumps work-in fact if you want to identify a "moving part" in diff pump, the oil is the closest thing it has to a moving part. Basically it gets heated, travels up the center of the pump stack, comes out in "jets" at the top(I think of this part as more or less working like a coffee percolator) and the falling drops from the oil jet "trap" residual air molecules. Simple conceptually and works well, they're just generally slower(slower to reach vacuum, can have great pumping capacity but that comes with a trade-off in how long it takes ones to heat up, can backstream oil if vented before cooling). On Agilent instruments, they have a heater and two temperature sensors-a "hot" sensor and a "too hot" sensor. The cooling fan also ties into the system, since the pump depends on the walls being cool and supposedly it will throw an error if the fan doesn't run.

In any case, even though even in careful use they'll often still shed a few drops of oil, it's pretty unusual to find one that's low. It would have needed to be turned upside down completely most likely, or perhaps vented while still hot repeatedly. I'd dread using the MS after either of those happened. There are probably 30 year old MSs out there that have never had the diff pump oil even looked at. Still, though, on the 73/75 it's easy to check-remove the lower cover and there's a sight glass on the front, behind the fan, that you can look at.

[WahWash!]

...

Thanks for the response! See I was actually leaning towards it being the main board - the firmware updated fine, heater controls and monitoring are functioning, and it's talking to chemstation so I figured it was some sort of corruption or fault in the OS, which seems to live on the EEPROMs behind the Log Amp.

My biggest fear is replacing the wrong board and doubling the expense of the the repair

I really wish there was someone out there that could re-flash those chips to see if that brought it back online...

I guess it could be one of the temp sensors on the diffusion pump but I'm not sure what resistance I should expect to see in a test - While the service manual gave test locations for voltage rails (and all of them checked out) I wasn't able to find anything that let me test deeper.

[benhutcherson]

...

Thanks for the response! See I was actually leaning towards it being the main board - the firmware updated fine, heater controls and monitoring are functioning, and it's talking to chemstation so I figured it was some sort of corruption or fault in the OS, which seems to live on the EEPROMs behind the Log Amp.

My biggest fear is replacing the wrong board and doubling the expense of the the repair

I really wish there was someone out there that could re-flash those chips to see if that brought it back online...

I guess it could be one of the temp sensors on the diffusion pump but I'm not sure what resistance I should expect to see in a test - While the service manual gave test locations for voltage rails (and all of them checked out) I wasn't able to find anything that let me test deeper.

There are some service out there that repair boards. You'll likely find the names mentioned here, although I don't recall them off the top of my head. You're probably looking at a couple hundred dollars to have a board repaired, but it beats a thousand+ for a replacement.

It might be worth reaching out and seeing if they'd have any advice about narrowing down which board, or if nothing else maybe ship them both the mainboard and the smartcard and see if they can diagnose and just repair the bad one.

BTW, for reference, I just had a mainboard changed in my 73. I got it from Resolution Labs in Indiana(outside West Lafayette). Their initial quote was $1295 for the board with $400 trade-in for old board. The people actually paying this bill wanted to have them come and install it(I could have easily done it myself-I had the old out a few times in diagnostics) for warranty reasons. It ended up being around $3500 to travel to St. Louis, but they billed the board at the same price.

This was for the earliest model mainboard-again HP branded instrument that started life as HPIB. The newer revision mainboard might be more. AFAIK the Smartcard should be the same as the one in my instrument-I'm not sure if there have been any revisions of the SCIII.

[WahWash!]

Awesome, thanks a bunch for all the help! We'll reach out to see if they've seen this before and can tell conclusively which board is the problem and get some quotes...

It's a real tossup between a repair and a $750 ebay mainboard I've been eyeing up. While the ebay one does have a warranty and can be returned if it doesn't work - it could add to the confusion if it has some separate issue and gives different symptoms.

Did get a quote for ~1200 for repair from a place in California, but I'm looking at weeks without the instrument if I go that route - Where I have a possibility of a 30 minute fix if that eBay board solves it.

I've never been much of a gambler

[benhutcherson]

Awesome, thanks a bunch for all the help! We'll reach out to see if they've seen this before and can tell conclusively which board is the problem and get some quotes...

It's a real tossup between a repair and a $750 ebay mainboard I've been eyeing up. While the ebay one does have a warranty and can be returned if it doesn't work - it could add to the confusion if it has some separate issue and gives different symptoms.

Did get a quote for ~1200 for repair from a place in California, but I'm looking at weeks without the instrument if I go that route - Where I have a possibility of a 30 minute fix if that eBay board solves it.

I've never been much of a gambler

Just make sure the Ebay board is the correct PN, especially since that sounds cheap to me for a later board.

I'm assuming you have a 5973N or later given that you mention the front panel.

[James_Ball]

We have on 5973 that was upgraded from HPIB to Lan with slow electronics that thinks the diff pump is a turbo on the manual tune screen, while on the pumpdown screen it will call it a diffusion pump. Of course that is also running E.02.01 software on Windows 11 so how knows where that bug comes from. It will also show the diff pump temperature as cold until about the next day even though it has already come to temp within less than an hour.

The log amp mismatch is definitely it thinking it has one type of side board while it actually has another, either slow or fast just not sure which it thinks you have versus which it actually has. I don't know all the part numbers for fast and slow main and side boards but I am sure there is a list somewhere. Just need to make sure they are a matched set.

[benhutcherson]

The log amp mismatch is definitely it thinking it has one type of side board while it actually has another, either slow or fast just not sure which it thinks you have versus which it actually has. I don't know all the part numbers for fast and slow main and side boards but I am sure there is a list somewhere. Just need to make sure they are a matched set.

Probably not a hard and fast rule, but most of the slow electronics PNs start with G1099.

Of course too some G1099 PNs are still around. G1099-60435 is the stupidly expensive turbo power cable that's still used on 77 AFAIK(and I've mentioned this a few times, but if anyone ever needs one, I can give you Digikey PNs to build them for about $25).

G1099-65001 is the slow log amp. G3170-65001(apparently Agilent likes that suffix for the log amp) is the fast one, which is also listed for the 5975 and 5977.

G1099-65015 is the slow sideboard. G3169-65015 is the fast one.

I seemingly was wrong about there being different mainboards. The only PN I've found is G1099-65010. I've also found a reference to a G1099-65010, but I can't seem to find any difference between the two.

The only thought I'd MAYBE have is does anyone know what part(s) change between the unipolar and bipolar emod? Something obviously has to be different, although I'd guess maybe it's in the PSU board and/or toroidal transformer?